Aim: To establish a linear additive model for the predication of in vitro sinomenine hydrochloride release from the combination of immediate release, enteric-coated and sustained-release pellets based on the release profiles of each pellet type. Methods: Immediate release pellets were manufactured by extrusion/spher-onization technology. The operation of bottom-spraying in the fluid-bed equipment was conducted to enteric-coating using Eudragit~(R) L-30D-55 and sustained-release coating using Surelease~(R) . In vitro sinomenine hydrochloride release profiles of both uncoated and coated pellets were fitted to the chosen mathematical equations offered by the curve fitting toolbox of Matlab~(R) before a linear additive model was created based upon the best-to-fitting equations. The proportion of each pellet type in the combined format to generate the desired 24 h sinomenine hydrochloride release profile was solved by Matlab~(R). The predicted and assayed sinomenine hydrochloride release from the polled pellets was compared. Results: It was shown that the actual sinomenine hydrochloride release profiles of each pellet type were approximate to those of predicted ones. A linear additive model of the appropriate mathematical equations of each pellet was proven to be capable of controlling in vitro release of sinomenine hydrochloride multiple-unit pellets. Conclusion: A multiple-unit combined system of the selected pellets, as a novel sustained-release system, was successfully prepared. In vitro release performance of the calculated combination of each pellet type could be guaranteed by this approach in designing sustained-release drug delivery system.